Mercurial > repos > dereeper > roary_plots

comparison Roary/t/Bio/Roary/AssemblyStatistics.t @ 0:c47a5f61bc9f draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author dereeper
date Fri, 14 May 2021 20:27:06 +0000
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:c47a5f61bc9f
1 #!/usr/bin/env perl
2 use strict;
3 use warnings;
4 use Data::Dumper;
5 use Test::Files;
6
7 BEGIN { unshift( @INC, './lib' ) }
8
9 BEGIN {
10 use Test::Most;
11 use_ok('Bio::Roary::AssemblyStatistics');
12 }
13
14 my $obj;
15 ok( $obj = Bio::Roary::AssemblyStatistics->new( spreadsheet => 't/data/input_block_spreadsheet.csv' ), 'initialise spreadsheet' );
16
17 my @genes = sort keys %{ $obj->_genes_to_rows };
18 is_deeply(
19 \@genes,
20 [
21 'SBOV29371', 'SBOV38871', 'SBOV43201', 'STY3593', 'STY4162', 'bcsC_1', 'betC_2', 'comM_2',
22 'dmsA4_1', 'dosC', 'dsbA_3', 'fadH_1', 'fimD_3', 'fliB_2', 'fliF', 'ftsN',
23 'gatY_1', 'glfT2', 'group_1000', 'group_1001', 'group_1004', 'group_1006', 'group_1009', 'group_220',
24 'group_277', 'group_281', 'group_283', 'group_284', 'group_288', 'hemD', 'hsrA_2', 'icsA',
25 'kdpD', 'ligB_1', 'marT_1', 'nepI', 'rffH', 'rpoS', 'selA_1', 'speC_3',
26 'sptP', 'srgB', 'stp', 'tmcA', 'tub', 'yadA', 'ybbW_1', 'yhaO_2',
27 'yicJ_1', 'yigZ'
28 ],
29 'all gene rows available'
30 );
31
32 is_deeply(
33 $obj->ordered_genes,
34 [
35 'dmsA4_1', 'group_1000', 'group_1001', 'SBOV43201', 'dosC', 'stp', 'fliB_2', 'fliF',
36 'dsbA_3', 'srgB', 'fimD_3', 'betC_2', 'tmcA', 'tub', 'rffH', 'hemD',
37 'group_1006', 'STY3593', 'group_1004', 'yigZ', 'group_220', 'glfT2', 'kdpD', 'speC_3',
38 'ybbW_1', 'sptP', 'SBOV29371', 'rpoS', 'fadH_1', 'yhaO_2', 'bcsC_1', 'STY4162',
39 'yadA', 'ligB_1', 'icsA', 'marT_1', 'selA_1', 'nepI', 'gatY_1', 'SBOV38871',
40 'group_288', 'hsrA_2', 'group_281', 'group_283', 'group_284', 'yicJ_1', 'ftsN', 'group_277',
41 'group_1009', 'comM_2'
42 ],
43 'ordered genes'
44 );
45
46 is_deeply(
47 $obj->sample_names_to_column_index,
48 {
49 'threeblocks' => 18,
50 'nocontigs' => 17,
51 'contigwithgaps' => 16,
52 'oneblock' => 14,
53 'threeblocksinversion' => 19,
54 'oneblockrev' => 15
55 },
56 'sample names to column index'
57 );
58
59 is_deeply( $obj->_sample_statistics('oneblock'), { num_blocks => 1, largest_block_size => 50 }, 'one block' );
60 is_deeply( $obj->_sample_statistics('oneblockrev'), { num_blocks => 1, largest_block_size => 50 }, 'one block reversed' );
61 is_deeply(
62 $obj->_sample_statistics('contigwithgaps'),
63 { num_blocks => 1, largest_block_size => 50 },
64 'one block where there are gaps everywhere'
65 );
66 is_deeply( $obj->_sample_statistics('nocontigs'), { num_blocks => 50, largest_block_size => 1 }, 'no contiguous blocks' );
67 is_deeply( $obj->_sample_statistics('threeblocks'), { num_blocks => 3, largest_block_size => 21 }, 'three blocks' );
68 is_deeply(
69 $obj->_sample_statistics('threeblocksinversion'),
70 { num_blocks => 3, largest_block_size => 20 },
71 'three blocks with an inversion in the middle'
72 );
73 is_deeply( $obj->gene_category_count, { core => 50 }, 'Gene category counts' );
74
75 # t/data/gene_category_count.csv
76 ok( $obj = Bio::Roary::AssemblyStatistics->new( spreadsheet => 't/data/gene_category_count.csv' ),
77 'initialise spreadsheet with variable numbers of genes in samples' );
78 is_deeply(
79 $obj->gene_category_count,
80 {
81 'core' => 1,
82 'cloud' => 4,
83 'soft_core' => 1,
84 'shell' => 24
85 },
86 'Categories as expected'
87 );
88 ok($obj->create_summary_output, 'create output file');
89 compare_ok('summary_statistics.txt', 't/data/expected_summary_statistics.txt', 'summary statistics as expected');
90
91
92 # t/data/gene_category_count.csv
93 ok( $obj = Bio::Roary::AssemblyStatistics->new( spreadsheet => 't/data/gene_category_count.csv', core_definition => 0.9667 ),
94 'initialise spreadsheet with core of 96.67%' );
95 is_deeply(
96 $obj->gene_category_count,
97 {
98 'core' => 1,
99 'soft_core' => 1,
100 'cloud' => 4,
101 'shell' => 24
102 },
103 'Categories as expected with cd of 96.67%'
104 );
105
106 # t/data/gene_category_count.csv
107 ok( $obj = Bio::Roary::AssemblyStatistics->new( spreadsheet => 't/data/gene_category_count.csv', core_definition => 0.9666 ),
108 'initialise spreadsheet with core of 96.66%' );
109 is_deeply(
110 $obj->gene_category_count,
111 {
112 'core' => 2,
113 'cloud' => 4,
114 'shell' => 24
115 },
116 'Categories as expected with cd of 96.66%'
117 );
118
119
120 unlink('summary_statistics.txt');
121 done_testing();